Climate Nexus developed an all-in-one dashboard where partners can easily access Climate Nexus polls and seamlessly browse survey data, create charts, suggest questions, map data, create PowerPoints, and more.
See tips below). The most recent polls are at the top🚨 If you’re not sure of the poll number, but know the topic or key word select ‘Search All Polls’ and in STEP 2, press DELETE/BACKSPACE and search the keyword(s) across all polls. After you select a question under ‘Search All Polls’ to identify the poll number note the text at the end of the question text.
🚨 Naming Convention when Searching Across All Polls: Q1x1… Do you have a favorable or unfavorable opinion of solar energy…pr1928 The first characters are the question number, and the last is the poll number. In between the ...s is the question text
🚨 Pro tip: Once you select the poll number, head on over to the top lines tab to browse all of the published questions for the poll
🚨 If you select a question with an x in the numbering you can look across the matrix. For example if we select question number Q1x1 in poll pr1928 and want to look across all of the energy sources to see which source of energy has the highest favorability rating you can click on the link that’s labeled Message Comparison Chart. There you can chart the relative favorability rating across the different energy sources.
Click on the relevant documentation tab below
After going through STEPS 1:3 in the Analysis tab, users have the ability download the charts they created in various formats by clicking the violin drop down menu at the top right of the chart. Users can also interact with the charts to add/remove answer options at the bottom of the charts.
Available export options:
Climate Nexus may or may not include top lines and cross tabs. To access a poll’s toplines and/or cross tabs you must select a poll number in STEP 1. Then click the Toplines or the Crosstabs tab to the right of the Analysis tab.
🚨 Users may copy and paste the cross tab data into excel. There are currently no options to download toplines and cross tabs.
Nexus Polling uses scientific online polling in conjunction with multilevel regression and synthetic post-stratification (MRsP)1 to calculate estimates for each state, CD, and city in the country. Multilevel regression and post-stratification is widely used to produce reliable and accurate estimates in small geographic areas without the need to conduct individual polls within each district - a methodology that emerged from political science and has been called the “gold standard” for estimating sub-national opinion (Selb and Munzert 2011, 456)
To map data down to the state, CD or city level:
(MRP) in its name. Only polls with (MRP) can utilize question mapping.State Maps CD maps City Metro Maps)🚨 If there is no MRP data for the question you’re looking for contact James Wyatt via email
🚨 If you want to quickly search a particular geographic unit, there is a data table at the bottom with a search function
🚨 Maps can be downloaded and embedded in your work streams. A full HTML file with the data is provided when you download
MRP works by combining information from large national survey samples, census data, demographic characteristics, and political data to accurately measure an area’s public opinion. Nexus Polling utilizes an extension of traditional MRP, called MRsP, in combination with hierarchical modeling to develop state estimates. This approach has been shown to increase the prediction and precision of sub-national public opinion estimation beyond traditional MRP by using synthetic joint distributions that are created on the marginal distributions. In comparison, traditional MRP utilizes “true” joint distributions, or nested stratas that are typically found in the Census, such as the interaction of age + gender + race, and is therefore limited to few variables. As an alternative, we use a technique called multidimensional iterative proportional fitting (mipfp) to develop cell proportions for each ‘type’ of person in a given state by calculating the joint probability distributions, based on the marginal distributions of known demographic variables, which is an exercise in spatial microsimulation.
Climate Nexus developed a prediction matrix and joint probability distribution table based on the Census marginal proportions of age, gender, race, household income, educational attainment, housing tenure occupancy (owner/renter) and Hispanic ethnicity resulting in 2,400 ‘types’ of individuals, or ‘demographic categories’. (for example, calculating the proportion of men in Florida age 65+ that are white with a household income between $50,000 - $100,000, have a bachelor’s degree owns a home and is of Hispanic ethnicity… and so on). By sidestepping the stringent data requirements of traditional MRP, we are able to develop dynamic and robust predictive models that include more predictive variables to better assess public opinion within small geographic boundaries.
\[N_{stratas} = 2(gender) * 3(race) * 4(education) * 2(Hispanic) * 5(household income) * 5(age) * 2(owner/renter). = 2,400\]
After the proportions of each ‘type’ of individual is calculated, we use a hierarchical Bayesian regression model to develop the predicted estimates for each cell row, and then we post stratify to take the weighted sum across all demographic categories (sum the vector) to make inferences about each state.
We use the national survey data to fit a generalized mixed effects model:
\[\widehat{y}=Pr(y_i=1) = logit^{-1}(\alpha_{j[i]} + \beta x_i + \epsilon_{i})\]
where
\[ logit^{-1}(\alpha) = \frac{exp(\alpha)}{exp(\alpha) + 1}\]
where distributions drawn with mean zero and estimated variance (below are the individual-level predictive variables - random effects variables):
\[ \begin{aligned} \alpha ^{race}_{j} \sim N(0, \sigma ^{2}_{race}), for j= 1,...,3 \\ \alpha ^{age}_{k} \sim N(0, \sigma ^{2}_{age}), for k= 1,...,5 \\ \alpha ^{education}_{l} \sim N(0, \sigma ^{2}_{education}), for l= 1,...,4 \\ \alpha ^{gender}_{m} \sim N(0, \sigma ^{2}_{gender}), for m= 1,...,2 \\ \alpha ^{own_rent}_{or} \sim N(0, \sigma ^{2}_{own_rent}), for or= 1,...,2 \\ \alpha ^{state}_{d} \sim N(0, \sigma ^{2}_{state}), for d= 1,...,50 \\ \alpha ^{Hispanic}_{h} \sim N(0, \sigma ^{2}_{Hispanic}), for h= 1,...,2 \\ \alpha ^{hhincome}_{hhi} \sim N(0, \sigma ^{2}_{hhincome}), for hhi= 1,...,5 \\ \end{aligned} \]
In the model, an individual’s response to a particular survey question (the outcome variable) is a function of their individual level demographic variables (random effects with varying intercepts), state grouping variables (non random effects), and interactions. In the case of gas use in the home and other particular questions in the pr1928 survey, traditional predictive variables like political party identification and educational attainment are much less predictive than say region/state, percent that live in urban/suburban, owner/renter, etc. As a result, the statistical models also include housing characteristics to serve as state-level grouping variables such as heating fuel (percent of homes that have utility gas hookups in a state), urban/rural percentages, and housing types. For state-grouping level covariates we also included 2016 election returns to inform the statistical models, which is a strong predictor of state aggregated public opinion.
For individual \(i\), with the following indexes for each demographic variable in the model, the model can be defined:
\[\begin{aligned} \widehat{y}=Pr(y_i=1) = logit^{-1}(\beta_0 + \alpha^{race}_{j[i]} + \alpha^{age}_{k[i]} + \alpha^{education}_{l[i]} + \alpha^{state}_{d[i]} + + \alpha^{ownerRenter}_{or[i]} \alpha^{gender}_{m[i]} + \\ \alpha^{Hispanic}_{h[i]} + \alpha^{hhincome}_{hhi[i]} + \alpha election_{2016} + \alpha CensusHouseHeatingFuel{5yearestimate} + \alpha PercentStateUrban \end{aligned} \]
The prediction for each strata, that is, the prediction for each “type of person” is weighted by the population frequency of the cell.
To develop cross tabs for particular a demographic group, we first calculate the proportion of that demographic type in a given geographic region, we then filter for the specific subgroup (different levels within categories such as age, race, etc.) and develop new post-stratification weights by taking the cell weighted proportion and dividing it by the sum of the weights of that subgroup. In the final step, we sum the post-stratified predicted probabilities for each subgroup. If a demographic type does not have valid Census proportions (such as proportions of self identified Republicans in a given state, which isn’t tracked by the Census) we first develop an MRsP model to calculate the proportion of individuals that fit that demographic type, much like predicting any other survey question or outcome variable. We then use those proportions as given population percentages in a given state and can then filter, divide the sum of the weights of that variable, post-stratify, and sum across each cell.
We follow the process below to estimate opinions of a given subgroup.
\[m = poststratification weights of sub-population k\] \[ \begin{aligned} \widehat{y}=Pr(y_i=1) = logit^{-1}(\alpha_{j[i]} + \beta x_i + \epsilon_{i}) \\ logit^{-1} = \frac{exp(a)}{1 + exp(a)} \\ \widehat{weightedStatePred}= \sum_{k=1}^{n} logit^{-1}(\alpha_{j[i]} + \beta x_i + \epsilon_{i}) * \frac{m_i}{\sum m} \end{aligned} \]
To develop final estimates and error associated with the statistical models, we utilize statistical bootstrapping. Bootstrapping allows us to check the stability of the results.
To estimate the overall margin of error for each state, we run bootstrap samples for each answer option in the survey and calculate the confidence interval at the 95% confidence level. We then take the mean of all confidence intervals for a given state as the final margin of error.
Because each estimate (for a specific answer choice in the survey) is a modeled outcome variable with its own corresponding confidence interval, questions may or may not add up to 100%. If they do not, we normalize results to help with interpretation.
TBD
How often is the survey updated?
To see when the portal was last updated, check the left panel.
Can I share information that I download or see on the portal?
Our partners are encouraged to share data internally with their organization, but the data (charts, maps, top lines, cross tabs, etc.) cannot legally be shared externally without the consent of Climate Nexus. We ask that you not share data with the media without consent. If you wish to use our data for presentations and other documents, we request that you contact us prior.
Can we share logins?
This isn’t advisable. Contact James Wyatt to request your own login.
How was this dashboard built?
The dashboard was built using a combination of R, Javascript, API backends + servers, relevant Shiny-R packages.